RU2766560C2 - Method and device for decoding ambiophonic audio presentation of sound field for audio playback using 2d-arrangements - Google Patents

Abstract

FIELD: audio processing.

SUBSTANCE: invention relates to means for processing audio presentation of a sound field. An audio signal is processed in an ambiophonic format in presentation of L speakers, based on a processing matrix. The processing matrix has elements based on speaker positions, while the processing matrix is determined based on weighing at least one element of the first matrix by means of a weight factor

. The first matrix is determined based on L speaker positions and at least a virtual position of at least a virtual speaker that is added to L speaker positions.

EFFECT: increase in the efficiency of processing audio presentation of the sound field.

3 cl, 7 dwg

Description

FIELD OF TECHNOLOGY TO WHICH THE INVENTION RELATES

The present invention relates to a method and apparatus for decoding an audio sound field representation, and in particular an ambiophonically formatted audio representation, for audio playback using a two-dimensional or near two-dimensional layout.

BACKGROUND OF THE INVENTION

Accurate localization is a key goal for any spatial audio playback system. Such playback systems are useful for conferencing, gaming, or other virtual environments that benefit from 3D audio. Soundstages in 3D can be synthesized or captured as a natural sound field. Sound field signals, such as, for example, ambiphony, carry the representation of the desired sound field. The decoding process is required to obtain individual speaker signals from the sound field representation. Decoding an ambiophonically formatted signal is also referred to as "processing". In order to synthesize audio scenes, panning functions that refer to the spatial layout of the speakers are required to obtain the spatial localization of a given sound source. To record a natural sound field, microphone arrays are required to capture spatial information. The ambiophonic approach is a very suitable tool for doing this. Ambiophonically formatted signals transfer the representation of the desired sound field based on the spherical harmonic decomposition of the sound field. While the basic ambiophonic format or B-format uses zeroth and first order spherical harmonics, the so-called high order ambiphony (HOA) also additionally uses at least 2nd order spherical harmonics. The spatial arrangement of the speakers is called the speaker layout. The decoding process requires a decoding matrix (also called a processing matrix) that is specific to some given speaker layout and that is generated using known speaker positions.

Commonly used speaker layouts are the stereo layout that uses two speakers, the standard surround layout that uses five speakers, and the surround extensions that use more than five speakers. However, these well-known arrangements are limited to two dimensions (2D), for example, no height information is reproduced. Processing for known speaker layouts that can reproduce height information has disadvantages in sound localization and coloration: either spatial vertical panoramas are perceived with very uneven loudness, or loudspeaker signals have strong sidelobes, which are disadvantageous especially for off-center listening positions. Thus, the so-called power-saving processing construction is preferable when processing the HOA sound field description for loudspeakers. This means that the processing of one sound source results in loudspeaker signals with constant energy regardless of the direction of the source. In other words, the input energy carried by the ambiophonic representation is stored by the loudspeaker processing means. International Patent Publication WO2014/012945A1 [1] of the present inventors describes the construction of a HOA processing tool with good power saving and localization properties for 3D speaker layouts. However, while this approach works well for 3D speaker layouts that cover all directions, some source directions are attenuated for 2D speaker layouts (such as 5.1 surround sound). This is especially true for directions where no loudspeakers are placed, such as from above.

In the work of F. Zotter and M. Frank "All-Round Ambisonic Panning and Decoding" [2], an "imaginary" loudspeaker is added if there is a void in the convex hull formed by the loudspeakers. However, the resulting signal for this imaginary loudspeaker is passed through the actual loudspeaker in playback. Thus, a signal source from that direction (ie the direction where no actual loudspeaker is located) will still be attenuated. In addition, this work shows the use of an imaginary loudspeaker only for use with VBAP (vector-based amplitude panning).

SUMMARY OF THE INVENTION

Thus, it remains a challenge to design energy-saving ambiophonic processing for 2D (two-dimensional) speaker layouts in which sound sources from directions where no speakers are placed are less attenuated or not attenuated at all. 2D speaker layouts can be classified as those in which the speaker elevation angles are within a certain small range (eg <10°) such that they are close to the horizontal plane.

This technical specification describes a solution for processing/decoding an ambiophonically formatted sound field audio representation for uniform or non-uniform spatial speaker distributions, the processing/decoding providing greatly improved localization and coloration properties and being energy saving, and even sound from directions in which no loudspeaker is available, being processed. Advantageously, sound from directions in which no speaker is available is processed with substantially the same energy and perceived loudness as it would have if the speaker were available in that direction. Of course, exact localization of these sound sources is not possible, since no loudspeaker is available in this direction.

In particular, at least some of the described embodiments provide a novel method for obtaining a decoding matrix for decoding sound field data in HOA format. Since at least the HOA format describes a sound field that is not directly related to speaker positions, and since the speaker signals to be received are necessarily in channel-based audio format, the decoding of HOA signals is always directly related to audio signal processing. As such, the same also applies to other sound field audio formats. Thus, the present disclosure relates to both decoding and processing of sound field-related audio formats. The terms "decoding matrix" and "processing matrix" are used interchangeably.

To obtain a decoding matrix for some given arrangement with good power saving properties, one or more virtual speakers are added at positions where no speaker is available. For example, to obtain an improved decoding matrix for a 2D layout, two virtual speakers are added at the top and bottom (corresponding to elevation angles of +90° and -90°, with 2D speakers placed at approximately 0° elevation). For this virtual 3D speaker layout, a decoding matrix is constructed that satisfies the energy saving property. Finally, the weights from the decoding matrix for the virtual speakers are mixed with constant gains for the actual speakers in the 2D layout.

According to one embodiment, a decoding matrix (or processing matrix) for processing or decoding an audio signal in ambiophonic format for some given set of speakers is generated by generating a first preliminary decoding matrix using a standard method and using modified speaker positions, where the modified speaker positions include the positions speakers of a given set of speakers and at least one additional virtual speaker position, and downmixing the first pre-decoding matrix, wherein the coefficients related to at least one additional virtual speaker are removed and distributed among the coefficients related to the speakers of a given set of speakers. In one embodiment, a subsequent decoding matrix normalization step follows this. The resulting decoding matrix is suitable for processing or decoding an ambiophonic signal for a given set of speakers, even sound from positions where no speaker is present is reproduced with the correct signal energy. This is due to the structure of the improved decoding matrix. Preferably, the first decoding pre-matrix is energy saving.

, где L является количеством громкоговорителей в 2D-компоновке громкоговорителей.In one embodiment, the decoding matrix has L rows and O3D columns. The number of rows corresponds to the number of speakers in the 2D speaker layout, and the number of columns corresponds to the number of ambiophonic coefficients O3D, which depends on the order N HOA according to O3D=(N+1)2. Each of the coefficients of the decoding matrix for the 2D speaker layout is the sum of at least the first intermediate coefficient and the second intermediate coefficient. The first intermediate coefficient is obtained by a power-saving 3D matrix construction method for the current speaker position of a 2D speaker layout, wherein the energy-saving 3D matrix construction method uses at least one virtual speaker position. The second intermediate coefficient is obtained by a coefficient which is obtained from said energy-saving 3D matrix construction method for at least one virtual speaker position multiplied by the weighting coefficient g. In one embodiment, the weighting factor g is calculated according to

where L is the number of speakers in the 2D speaker layout.

In one embodiment, the invention relates to a computer-readable storage medium having executable instructions stored thereon for causing a computer to execute a method comprising the steps of the method disclosed above or in the claims.

The device that operates the method is disclosed in clause 9.

Advantageous embodiments are disclosed in the dependent claims, in the following description and in the drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

Exemplary embodiments of the invention are described with reference to the accompanying drawings, which are shown in

1 is a flowchart of a method according to one embodiment;

2 exemplary structure of a downmixed HOA decoding matrix;

3 is a block diagram for obtaining and modifying speaker positions;

4 is a block diagram of a device according to one embodiment;

5 the energy distribution resulting from the standard decoding matrix;

Fig. 6 energy distribution resulting from the decoding matrix according to the embodiments; And

7 using separately optimized decoding matrices for different frequency bands.

DESCRIPTION OF EMBODIMENTS

1 ...

L для L громкоговорителей вводятся i10 в процесс. Следует заметить, что, когда упоминаются позиции, фактически здесь подразумеваются пространственные направления, т.е. позиции громкоговорителей определяются их углами наклона

и углами азимута

, которые комбинируются в вектор

. Затем по меньшей мере одна позиция виртуального громкоговорителя добавляется 10. В одном варианте осуществления все позиции громкоговорителей, которые вводятся в процесс i10, находятся по существу в одной и той же плоскости так, что они составляют 2D-компоновку, и по меньшей мере один виртуальный громкоговоритель, который добавляется, находится вне этой плоскости. В одном особенно выгодном варианте осуществления все позиции громкоговорителей, которые вводятся в процесс i10, находятся по существу в одной и той же плоскости, и позиции двух виртуальных громкоговорителей добавляются на этапе 10. Имеющие преимущества позиции двух виртуальных громкоговорителей описаны ниже. В одном варианте осуществления добавление выполняется согласно ур.(6) ниже. Этап 10 добавления дает в результате модифицированный набор углов громкоговорителей

'1 ...

'L+Lвирт на q10. Lвирт является количеством виртуальных громкоговорителей. Модифицированный набор углов громкоговорителей используется на этапе 11 построения 3D-матрицы декодирования. Также порядок N HOA (в общем случае порядок коэффициентов сигнала звукового поля) должен быть обеспечен i11 для этапа 11.1 shows a flowchart of a method for decoding an audio signal, in particular a sound field signal, according to one embodiment. Decoding sound field signals generally requires speaker positions for which the audio signal must be processed. These speaker positions

one ...

L for L speakers are introduced i10 into the process. It should be noted that when positions are mentioned, in fact, spatial directions are meant here, i.e. speaker positions are determined by their tilt angles

and azimuth angles

, which are combined into a vector

. Then, at least one virtual speaker position is added 10. In one embodiment, all speaker positions that are entered into the i10 process are substantially in the same plane so that they constitute a 2D layout, and at least one virtual speaker The , which is being added, is outside this plane. In one particularly advantageous embodiment, all speaker positions that are entered into process i10 are essentially in the same plane, and two virtual speaker positions are added in step 10. Beneficial two virtual speaker positions are described below. In one embodiment, the addition is performed according to Eq. (6) below. The adding step 10 results in a modified set of speaker angles

'one ...

'L+Lvirt on q10. Lvirt is the number of virtual speakers. The modified set of speaker angles is used in step 11 of building a 3D decoding matrix. Also the order N HOA (generally the order of the sound field signal coefficients) must be provided by i11 for step 11.

The 3D decoding matrix building step 11 executes any known method for generating a 3D decoding matrix. Preferably, the 3D decoding matrix is suitable for a power saving type of decoding/processing. For example, the method described in PCT/EP2013/065034 can be used. The 3D decoding matrix building step 11 results in a decoding matrix or processing matrix D' that is suitable for processing L'=L+Lvirt speaker signals, where Lvirt is the number of virtual speaker positions that were added in step 10 "add virtual speaker positions ".

, которая имеет L строк, т.е. меньше строк, чем матрица декодирования D', но имеет то же самое количество столбцов, что и матрица декодирования D'. Иными словами, размерность матрицы декодирования D' равна (L+Lвирт)×O3D и размерность микшированной с понижением 3D-матрицы декодирования

равна L×O3D.Since only L speakers are physically available, the decoding matrix D' that results from the 3D decoding matrix construction step 11 must be adapted to the L speakers in the downmix step 12 . This step performs a downmix of the decoding matrix D', where the coefficients related to the virtual speakers are weighted and distributed among the coefficients related to the existing speakers. Preferably, the coefficients of any particular HOA order (ie, a column of the decoding matrix D') are weighted and added to the coefficients of the same HOA order (ie, the same column of the decoding matrix D'). One example is the downmix according to Equation (8) below. The downmix step 12 results in a downmixed 3D decoding matrix

, which has L rows, i.e. fewer rows than the decoding matrix D' but has the same number of columns as the decoding matrix D'. In other words, the dimension of the decoding matrix D' is (L+Lvirt)×O3D and the dimension of the downmixed 3D decoding matrix

equals L×O3D.

из HOA-матрицы декодирования D'. HOA-матрица декодирования D' имеет L+2 строк, что означает, что две виртуальные позиции громкоговорителей были добавлены к L доступным позициям громкоговорителей, и O3D столбцов, где O3D=(N+1)2 и N является порядком HOA. На этапе 12 понижающего микширования коэффициенты строк L+1 и L+2 HOA-матрицы декодирования D' взвешиваются и распределяются по коэффициентам их соответственного столбца, и строки L+1 и L+2 удаляются. Например, первые коэффициенты d'L+1,1 и d'L+2,1 каждой из строк L+1 и L+2 взвешиваются и добавляются к первым коэффициентам всех остальных строк, таким как d'1,1. Получающийся в результате коэффициент

1,1 микшированной с понижением HOA-матрицы декодирования

является функцией от d'1,1, d'L+1,1, d'L+2,1 и весового коэффициента g. Тем же самым образом, например, получающийся в результате коэффициент

2,1 микшированной с понижением HOA-матрицы декодирования

является функцией от d'2,1, d'L+1,1, d'L+2,1 и весового коэффициента g, и получающийся в результате коэффициент

1,2 микшированной с понижением HOA-матрицы декодирования

является функцией от d'1,2, d'L+1,2, d'L+2,2 и весового коэффициента g.2 shows an exemplary structure of a downmixed HOA decoding matrix

from the HOA decoding matrix D'. The HOA decoding matrix D' has L+2 rows, meaning that two virtual speaker positions have been added to the L available speaker positions, and O3D columns, where O3D=(N+1)2 and N is the HOA order. In the downmix step 12, the coefficients of the rows L+1 and L+2 of the HOA decoding matrix D' are weighted and allocated to the coefficients of their respective column, and the rows L+1 and L+2 are removed. For example, the first coefficients d'L+1.1 and d'L+2.1 of each of the rows L+1 and L+2 are weighted and added to the first coefficients of all other rows, such as d'1,1. The resulting coefficient

1.1 HOA downmix decoding matrix

is a function of d'1,1, d'L+1,1, d'L+2,1 and the weighting factor g. In the same way, for example, the resulting coefficient

2.1 HOA downmix decoding matrix

is a function of d'2.1, d'L+1.1, d'L+2.1 and the weight factor g, and the resulting factor

1.2 HOA downmix decoding matrix

is a function of d'1,2, d'L+1,2, d'L+2,2 and the weighting factor g.

будет нормализована на этапе 13 нормализации. Однако этот этап 13 является опциональным, поскольку не нормализованная матрица декодирования также может быть использована для декодирования сигнала звукового поля. В одном варианте осуществления микшированная с понижением HOA-матрица декодирования

нормализуется согласно ур.(9) ниже. Этап 13 нормализации дает в результате нормализованную микшированную с понижением HOA-матрицу декодирования D, которая имеет ту же самую размерность L×O3D, что и микшированная с понижением HOA-матрица декодирования

.Typically downmixed HOA decoding matrix

will be normalized in step 13 normalization. However, this step 13 is optional since a non-normalized decoding matrix can also be used to decode the sound field signal. In one embodiment, the downmixed HOA decoding matrix

is normalized according to eq.(9) below. The normalization step 13 results in a normalized downmix HOA decoding matrix D which has the same dimension L×O3D as the downmixed HOA decoding matrix

.

The normalized downmix HOA decoding matrix D can then be used in the sound field decoding step 14 where the sound field input signal i14 is decoded for L speaker signals q14. Typically, the normalized downmix HOA decoding matrix D does not need to be modified as long as the speaker layout is not modified. Thus, in one embodiment, the normalized downmix HOA decoding matrix D is stored in a decoding matrix store.

1 ...

L L громкоговорителей и порядка N коэффициентов сигнала звукового поля, определения 102 из позиций, что L громкоговорителей находятся по существу в 2D-плоскости, и генерирования 103 по меньшей мере одной виртуальной позиции

виртуального громкоговорителя.3 depicts details of how, in one embodiment, speaker positions are obtained and modified. This embodiment contains the steps of determining 101 positions

one ...

LL speakers and order N sound field signal coefficients, determining 102 from positions that the L speakers are substantially in a 2D plane, and generating 103 at least one virtual position

virtual speaker.

является одной из

и

.In one embodiment, at least one virtual position

is one of

And

.

и

, соответствующие двум виртуальным громкоговорителям, генерируются 103, где

и

.In one embodiment, two virtual positions

And

, corresponding to two virtual speakers, are generated 103, where

And

.

1 ...

L L громкоговорителей и порядок N коэффициентов сигнала звукового поля, определяют 102 из позиций, что L громкоговорителей находятся по существу в 2D-плоскости, генерируют 103 по меньшей мере одну виртуальную позицию

виртуального громкоговорителя, генерируют 11 3D-матрицу декодирования D', причем определенные позиции

1 ...

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, микшируют с понижением 12 3D-матрицу декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и декодируют 14 закодированный аудиосигнал i14 с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов q14 громкоговорителей.According to one embodiment, a method for decoding an encoded audio signal for L speakers at known positions comprises the steps of determining 101 positions

one ...

Loudspeaker LL and order N of sound field signal coefficients, determine 102 from positions that L loudspeakers are substantially in a 2D plane, generate 103 at least one virtual position

virtual loudspeaker generate 11 3D decoding matrix D', with certain positions

one ...

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has coefficients for said defined and virtual speaker positions, the 3D decoding matrix D' is downmixed 12, wherein the coefficients for the virtual speaker positions are weighted and distributed over coefficients related to the defined speaker positions, and where a reduced 3D decoding matrix is obtained

having coefficients for certain speaker positions, and decode the 14 i14 encoded audio signal using the reduced 3D decoding matrix

, and a plurality of decoded speaker signals q14 are obtained.

In one embodiment, the encoded audio signal is a sound field signal, such as in HOA format.

виртуального громкоговорителя является одной из

и

.In one embodiment, at least one virtual position

virtual speaker is one of the

And

.

.In one embodiment, the coefficients for virtual speaker positions are weighted by a weighting factor

.

, причем получается нормализованная уменьшенная 3D-матрица декодирования D, и этап 14 декодирования закодированного аудиосигнала i14 использует нормализованную уменьшенную 3D-матрицу декодирования D. В одном варианте осуществления способ имеет дополнительный этап сохранения уменьшенной 3D-матрицы декодирования

или нормализованной микшированной с понижением HOA-матрицы декодирования D в хранилище матрицы декодирования.In one embodiment, the method has the additional step of normalizing the reduced 3D decoding matrix

, wherein a normalized reduced 3D decoding matrix D is obtained, and the encoded audio decoding step 14 i14 uses the normalized reduced 3D decoding matrix D. In one embodiment, the method has an additional step of storing the reduced 3D decoding matrix

or a normalized downmix HOA decoding matrix D in the decoding matrix store.

According to one embodiment, a decoding matrix for processing or decoding a sound field signal for a given set of speakers is generated by generating a first preliminary decoding matrix using a standard method and using modified speaker positions, where the modified speaker positions include the speaker positions of the given speaker set and at least one additional virtual speaker position, and downmixing the first pre-decoding matrix, wherein the coefficients associated with at least one additional virtual speaker are removed and distributed among the coefficients associated with speakers from a given set of speakers. In one embodiment, a subsequent decoding matrix normalization step follows this. The resulting decoding matrix is suitable for processing or decoding the sound field signal for a given set of speakers, even sound from positions where no speaker is present is reproduced with the correct signal energy. This is due to the structure of the improved decoding matrix. Preferably, the first decoding pre-matrix is energy saving.

1 ...

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, блок 412 понижающего микширования матрицы для понижающего микширования 3D-матрицы декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и блок 414 декодирования для декодирования закодированного аудиосигнала с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов громкоговорителей.4a) depicts a block diagram of a device according to one embodiment. A device 400 for decoding an encoded audio signal in sound field format for L speakers at known positions comprises a summing block 410 for adding at least one position of at least one virtual speaker to the L speaker positions, a decoding matrix generator block 411 for generating a 3D decoding matrix D ', and the positions

one ...

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has coefficients for said defined and virtual speaker positions, a matrix downmixer 412 for downmixing the 3D decoding matrix D', wherein the coefficients for the virtual speaker positions are weighted and allocated to the coefficients related to the defined positions loudspeakers, and moreover, a reduced 3D decoding matrix is obtained

, having coefficients for certain speaker positions, and a decoder 414 for decoding the encoded audio signal using a reduced 3D decoding matrix

, and a plurality of decoded speaker signals are obtained.

, причем получается нормализованная уменьшенная 3D-матрица декодирования D, и блок 414 декодирования использует нормализованную уменьшенную 3D-матрицу декодирования D.In one embodiment, the apparatus further comprises a normalizer 413 for normalizing the reduced 3D decoding matrix

, wherein a normalized reduced 3D decoding matrix D is obtained, and the decoding unit 414 uses the normalized reduced 3D decoding matrix D.

L) L громкоговорителей и порядка N коэффициентов сигнала звукового поля, второй блок 4102 определения для определения из позиций, что L громкоговорителей находятся по существу в 2D-плоскости, и блок 4103 генерирования позиции виртуального громкоговорителя для генерирования по меньшей мере одной виртуальной позиции (

) виртуального громкоговорителя.In one embodiment depicted in FIG. 4b), the device further comprises a first determiner 4101 for determining positions (

L) L loudspeakers and order N sound field signal coefficients, a second determiner 4102 to determine from positions that L loudspeakers are substantially in a 2D plane, and a virtual loudspeaker position generating block 4103 to generate at least one virtual position (

) virtual speaker.

In one embodiment, the apparatus further comprises a plurality of band pass filters 715b for separating the encoded audio signal into a plurality of frequency bands, wherein a plurality of separate 3D decoding matrices Db' are generated 711b, one for each frequency band, and each 3D decoding matrix Db' is downmixed 712b and optionally normalized separately, and wherein the decoder 714b decodes each frequency band separately. In this embodiment, the device further comprises a plurality of summing units 716b, one for each speaker. Each summation block sums the frequency bands that apply to the respective loudspeaker.

Each of summing block 410, decoding matrix generator block 411, matrix downmixing block 412, normalization block 413, decoding block 414, first determination block 4101, second determination block 4102, and virtual speaker position generation block 4103 may be implemented by one or more processors, and each of these blocks can share the same processor with any other of these or other blocks.

7 depicts an embodiment that uses separately optimized decoding matrices for different input signal bandwidths. In this embodiment, the decoding method comprises the step of dividing the encoded audio signal into a plurality of frequency bands using band pass filters. A plurality of separate 3D decoding matrices Db' are generated 711b, one for each frequency band, and each 3D decoding matrix Db' is downmixed 712b and optionally normalized separately. Decoding 714b of the encoded audio signal is performed for each frequency band separately. This has the advantage that frequency dependent differences in human perception can be taken into account and can result in different decoding matrices for different frequency bands. In one embodiment, only one or more (but not all) decoding matrices are generated by adding virtual speaker positions and then weighting and distributing their coefficients to the coefficients for the existing speaker positions, as described above. In another embodiment, each of the decoding matrices is generated by adding virtual speaker positions and then weighting and distributing their coefficients to the coefficients for the existing speaker positions, as described above. Finally, all frequency bands that belong to the same speaker are summed in one frequency band summation block 716b for each speaker, in the reverse operation of frequency band splitting.

Each of summing block 410, decoding matrix generator block 711b, matrix downmixing block 712b, normalization block 713b, decoding block 714b, frequency band summing block 716b, and bandpass filter block 715b may be implemented by one or more processors, and each of these blocks may be jointly use the same processor with any other of these or other blocks.

One aspect of the present disclosure is to provide a 2D layout processing matrix with good power saving properties. In one embodiment, two virtual speakers are added at the top and bottom (elevation angles of +90° and -90°, with the 2D speakers placed at approximately 0° elevation). For this virtual 3D speaker layout, a processing matrix is built that satisfies the power saving property. Finally, the weights from the processing matrix for the virtual speakers are mixed with constant gains for the actual speakers in the 2D layout.

The following describes ambiophonic (namely HOA) processing.

Ambiophonic processing is the process of computing loudspeaker signals from an ambiophonic description of the sound field. It is sometimes also called ambiophonic decoding. An ambiophonic 3D sound field representation of order N is considered, where the number of coefficients is

ZEqn1_O3d

ZEqn1_O3d (one)

c O3D элементами. Посредством матрицы обработки D

сигналы громкоговорителей для временной выборки t вычисляются следующим образомThe coefficients for time sample t are represented by the vector b(t)

with O3D elements. By processing matrix D

speaker signals for time sample t are calculated as follows

ZEqn1_HoaRendering

ZEqn1_HoaRendering (2)

, и w

, и L является количеством громкоговорителей.where D

, and w

, and L is the number of speakers.

и углы азимута

, которые комбинируются в вектор

для

. Различные расстояния громкоговорителей от позиций прослушивания компенсируются посредством отдельных задержек для каналов громкоговорителей.Loudspeaker positions are determined in terms of their tilt angles

and azimuth angles

, which are combined into a vector

for

. Different speaker distances from listening positions are compensated for by separate delays for the speaker channels.

The signal energy in the HOA region is given as follows

(3)

where H denotes the (complex conjugate) transposition. The corresponding energy of the loudspeaker signals is calculated as follows

(4)

для сберегающей энергию матрицы декодирования/обработки должно быть постоянным для того, чтобы достигать сберегающего энергию декодирования/обработки.Ratio

for the energy-saving decoding/processing matrix must be constant in order to achieve energy-saving decoding/processing.

As such, the following extension for enhanced 2D processing is proposed: To build processing matrices for 2D speaker layouts, one or more virtual speakers are added. 2D arrangements are understood to be those in which the elevation angles of the loudspeakers are within a certain small range so that they are close to the horizontal plane. This can be expressed as follows

ZEqn1_2dthreshold

ZEqn1_2dthreshold (five)

обычно выбирается для соответствия значению в диапазоне 5°-10° в одном варианте осуществления.Threshold

typically selected to match a value in the range of 5°-10° in one embodiment.

. Последние (в этом примере две) позиции громкоговорителей являются позициями двух виртуальных громкоговорителей на северном и южном полюсах (в вертикальном направлении, т.е. сверху и снизу) полярной системы координат:To construct the processing, a modified set of speaker angles is determined

. The last (two in this example) speaker positions are the positions of the two virtual speakers at the north and south poles (in the vertical direction, i.e. top and bottom) of the polar coordinate system:

ZEqn1_virtualLS

ZEqn1_virtualLS (6)

. Из этих модифицированных позиций громкоговорителей матрица обработки

строится с подходом сбережения энергии. Например, способ построения, описанный в [1], может быть использован. Теперь окончательная матрица обработки для исходной компоновки громкоговорителей находится из D'. Одной идеей является микшировать весовые коэффициенты для виртуального громкоговорителя, определенные в матрице D', для действительных громкоговорителей. Фиксированный коэффициент усиления используется, который выбирается равнымSo the new number of speakers used to build the processing is

. From these modified speaker positions, the processing matrix

built with an energy conservation approach. For example, the construction method described in [1] can be used. Now the final processing matrix for the original speaker layout is from D'. One idea is to mix the virtual speaker weights defined in the matrix D' to the actual speakers. A fixed gain is used, which is chosen to be

.ZEqn1_downmixGain

.ZEqn1_downmixGain (7)

(также называемой здесь уменьшенной 3D-матрицей декодирования) определяются следующим образомIntermediate matrix coefficients

(also referred to here as the reduced 3D decoding matrix) are defined as follows

для

и

ZEqn1_Ddownmix

for

And

ZEqn1_Ddownmix (8)

является элементом матрицы

в l-й строке и q-м столбце. На опциональном последнем этапе промежуточная матрица (уменьшенная 3D-матрица декодирования) нормализуется с использованием нормы Фробениуса:where

is an element of the matrix

in l-th row and q-th column. In an optional last step, the intermediate matrix (reduced 3D decoding matrix) is normalized using the Frobenius norm:

ZEqn1_matrixNorm

ZEqn1_matrixNorm (nine)

5 and 6 show the energy distributions for a 5.0 surround speaker layout. In both drawings, the energy values are shown as grayscale and the circles indicate the positions of the loudspeakers. By means of the disclosed method in particular, attenuation from above (and also from below, not shown here) is clearly reduced.

5 shows the energy distribution resulting from the standard decoding matrix. The small circles around the z=0 plane represent the positions of the loudspeakers. As can be seen, the energy range of [-3.9, ..., 2.1] dB is covered, resulting in an energy difference of 6 dB. In addition, signals from above (and below, not shown) the unit sphere are reproduced at very low energy, i.e. are inaudible as no loudspeakers are available here.

FIG. 6 depicts the energy distribution resulting from a decoding matrix according to one or more embodiments, with the same number of speakers in the same positions as in FIG. At least the following advantages are provided: firstly, a smaller energy range [-1.6, ..., 0.8] dB is covered, resulting in smaller energy differences of only 2.4 dB. Secondly, signals from all directions of the unit sphere are reproduced with their correct energy, even if no loudspeakers are available here. Since these signals are played through the available speakers, their localization is not correct, but the signals are heard at the correct volume. In this example, top and bottom signals (not shown) are made audible due to decoding by the enhanced decoding matrix.

1, ...,

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, микшируют с понижением 3D-матрицу декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и декодируют закодированный аудиосигнал с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов громкоговорителей.In one embodiment, a method for decoding an encoded audio signal in ambiophonic format for L speakers at known positions comprises adding at least one position of at least one virtual speaker to the L speaker positions, generating a 3D decoding matrix D', wherein the positions

one, ...,

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has the coefficients for said specific and virtual speaker positions, the 3D decoding matrix D' is downmixed, the coefficients for the virtual speaker positions are weighted and distributed over the coefficients related to the determined speaker positions, and the result is reduced 3D decoding matrix

having coefficients for certain speaker positions, and decode the encoded audio signal using the reduced 3D decoding matrix

, and a plurality of decoded speaker signals are obtained.

1 ...

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, блок 412 понижающего микширования матрицы для понижающего микширования 3D-матрицы декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и блок 414 декодирования для декодирования закодированного аудиосигнала с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов громкоговорителей.In another embodiment, a device for decoding an encoded audio signal in an ambiophonic format for L speakers at known positions comprises a summation block 410 for adding at least one position of at least one virtual speaker to the L speaker positions, a decoding matrix generator block 411 for generating a 3D matrix decoding D', and positions

one ...

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has coefficients for said defined and virtual speaker positions, a matrix downmixer 412 for downmixing the 3D decoding matrix D', wherein the coefficients for the virtual speaker positions are weighted and allocated to the coefficients related to the defined positions loudspeakers, and moreover, a reduced 3D decoding matrix is obtained

, having coefficients for certain speaker positions, and a decoder 414 for decoding the encoded audio signal using a reduced 3D decoding matrix

, and a plurality of decoded speaker signals are obtained.

1 ...

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, блок 412 понижающего микширования матрицы для понижающего микширования 3D-матрицы декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и блок 414 декодирования для декодирования закодированного аудиосигнала с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов громкоговорителей.In yet another embodiment, a device for decoding an encoded audio signal in ambiophonic format for L loudspeakers at known positions comprises at least one processor and at least one memory, the memory having stored instructions that, when executed in the processor, perform a summation block 410 for adding over of at least one position of at least one virtual speaker to the positions of L speakers, a decoding matrix generator block 411 for generating a 3D decoding matrix D', wherein the positions

one ...

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has coefficients for said defined and virtual speaker positions, a matrix downmixer 412 for downmixing the 3D decoding matrix D', wherein the coefficients for the virtual speaker positions are weighted and allocated to the coefficients related to the defined positions loudspeakers, and moreover, a reduced 3D decoding matrix is obtained

, having coefficients for certain speaker positions, and a decoder 414 for decoding the encoded audio signal using a reduced 3D decoding matrix

, and a plurality of decoded speaker signals are obtained.

1, ...,

L L громкоговорителей и по меньшей мере одна виртуальная позиция

используются, и 3D-матрица декодирования D' имеет коэффициенты для упомянутых определенных и виртуальных позиций громкоговорителей, микшируют с понижением 3D-матрицу декодирования D', причем коэффициенты для виртуальных позиций громкоговорителей взвешиваются и распределяются по коэффициентам, относящимся к определенным позициям громкоговорителей, и причем получается уменьшенная 3D-матрица декодирования

, имеющая коэффициенты для определенных позиций громкоговорителей, и декодируют закодированный аудиосигнал с использованием уменьшенной 3D-матрицы декодирования

, причем получается множество декодированных сигналов громкоговорителей. Дополнительные варианты осуществления машиночитаемых носителей информации могут включать в себя любые признаки, описанные выше, в частности признаки, раскрываемые в зависимых пунктах формулы изобретения, ссылающиеся на пункт 1.In yet another embodiment, a computer-readable storage medium has executable instructions stored thereon for causing a computer to perform a method for decoding an encoded audio signal in ambiophonic format for L speakers at known positions, the method comprising adding at least one position of at least one virtual speaker to L speaker positions generate a 3D decoding matrix D', with the positions

one, ...,

Loudspeaker LL and at least one virtual position

are used and the 3D decoding matrix D' has the coefficients for said specific and virtual speaker positions, the 3D decoding matrix D' is downmixed, the coefficients for the virtual speaker positions are weighted and distributed over the coefficients related to the determined speaker positions, and the result is reduced 3D decoding matrix

having coefficients for certain speaker positions, and decode the encoded audio signal using the reduced 3D decoding matrix

, and a plurality of decoded speaker signals are obtained. Additional embodiments of computer-readable storage media may include any of the features described above, in particular the features disclosed in dependent claims referring to claim 1.

It should be understood that the present invention has been described by way of example only, and modifications of the details may be made without departing from the scope of the invention. For example, although only described in relation to HOA, the invention can also be applied to other sound field audio formats.

Each feature disclosed in the description, (where appropriate) in the claims and drawings may be provided independently or in any appropriate combination. The features may, where appropriate, be implemented in hardware, software, or a combination of both. Reference numerals appearing in the claims are for illustrative purposes only and are not intended to limit the scope of the claims.

The following materials have been referenced above.

[1] International Patent Publication No. WO2014/012945A1 (PD120032).

[2] F. Zotter and M. Frank, "All-Round Ambisonic Panning and Decoding", Journal of the Society of Audio Engineers, 2012, Volume 60, pp. 807-820.

Claims (9)

1. A method for processing an audio signal in an ambiophonic format into a 2D speaker layout, comprising steps in which processing an audio signal in ambiophonic format into a representation of L loudspeakers based on a processing matrix, wherein the processing matrix has elements based on speaker positions, and wherein the processing matrix is determined based on weighting at least an element of the first matrix by means of a weighting factor

, And wherein the first matrix is determined based on the L speaker positions and at least the virtual position of at least the virtual speaker that is added to the L speaker positions. 2. A device for processing an audio signal in an ambiophonic format into a 2D loudspeaker layout, comprising: a handler for processing an audio signal in ambiophonic format into a representation of L loudspeakers based on a processing matrix, wherein the processing matrix has elements based on speaker positions, and wherein the processing matrix is determined based on weighting at least an element of the first matrix by means of a weighting factor

, And wherein the first matrix is determined based on the L speaker positions and at least the virtual position of at least the virtual speaker that is added to the L speaker positions. 3. A durable storage medium that contains or stores or has a digital audio signal recorded on it, decoded in accordance with paragraph 1.

Images